Lewis Acid Catalyzed Reactions of Chiral Imidazolidinones and

ARTICLE pubs.acs.org/joc

Lewis Acid Catalyzed Reactions of Chiral Imidazolidinones and Oxazolidinones: Insights on the Role of the Catalyst Filipe J. S. Duarte,† Snezhana M. Bakalova,†,‡ Eurico J. Cabrita,† and A. Gil Santos†,* †

REQUIMTE, CQFB, Departamento de Química, Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa, 2829-516 Caparica, Portugal ‡ Institute of Organic Chemistry, Bulgarian Academy of Sciences, Acad. G. Bonchev str., block 9, 1113 Sofia, Bulgaria

bS Supporting Information ABSTRACT: The mechanism proposed by Evans to justify the selectivity obtained in Lewis acid catalyzed DielsAlder reactions of cyclopentadiene with acyloxazolidinones has been generalized and used in the rationalization of selectivities obtained in many other systems. However, we recently proposed an alternative mechanism, on the basis of open-chain mono- and bicomplexes, that avoids the need for chelates and explains the selectivity obtained by Evans. In this manuscript we apply our proposal to the catalyzed conjugated addition of amines to acylimidazolidinones, reported by Cardillo, and we clearly show that aluminum chelates are not involved in the reaction, as they induce no selectivity, while Cardillo observed high experimental selectivities. Our data equally show that bicomplexes with carbonyl parallel orientation, proposed by Cardillo to justify the experimental selectivity with nonchelating Lewis acids, indeed induce the opposite selectivity and have also to be dismissed. On the other hand, our mechanistic proposal allows for the full rationalization of the data obtained by Cardillo with aluminum, boron, or zinc Lewis acids and supports our previous proposal on DA cycloadditions of dienes to Evans chiral auxiliary derivatives.

’ INTRODUCTION Chiral R,β-unsaturated N-acylimidazolidinones (1) and N-acyloxazolidinones (2) undergo a variety of Lewis acid catalyzed addition reactions with moderate to excellent asymmetric induction.1,2

While the enhanced electrophilicity of the β-carbon may be rationalized in a straightforward way as due to carbonyl complexation with the Lewis acid, the same is not true for the observed selectivities, considering the steric hindrance of the ring substituents (Scheme 1). The unexpected results obtained by Evans in DielsAlder (DA) additions of cyclopentadiene to oxazolidinones of type 2 (Scheme 1) were rationalized on the basis of a mechanism where the stereochemistry is attributed to a reacting conformer in which the two carbonyl groups are in coplanar orientation, as a result of their chelation with the LA (Scheme 1).1g Although the authors observed that LAs for which bidendate complexations are well established, e.g., TiCl4,3 were not effective catalysts, the concept r 2011 American Chemical Society

of bidendate chelation, contributing as an important organizational constraint to the transition structures (TSs), has been widely accepted and used thereafter to explain the diastereoselectivities in DA and related reactions.2a,4 However, the high energy needed for the amide-bond rotation5 and the observation that some systems originate the Evans stereochemistry without any kind of chelation to account for the coplanarity of the carbonyl groups led to the proposal of alternative reaction mechanisms6 that culminated in a recent proposal made by us for the rationalization of the stereoselectivity in dynamic kinetic resolution reactions (DKR).5,7 More recently, we also proposed an alternative mechanism for the LA promoted DielsAlder reaction of Evans auxiliary derivatives8 that explains the observed selectivities by the catalysis via low energetic open-chain mono- or bicomplexes, at the chain and the ring carbonyl groups, which are easily observed by NMR measurements (Scheme 2). As opposed to the Evans model, our proposal relies on a chirality transfer concept, in which an achiral Lewis acid works as a bridge for the transfer of chirality between the chiral auxiliary and the prochiral reactive center (Scheme 2). Thus, the preferential attack at the CR-si face is a result of the bulkiness of the LA, which strongly hinders the

Received: April 15, 2011 Published: July 26, 2011 6997

dx.doi.org/10.1021/jo201318f | J. Org. Chem. 2011, 76, 6997–7004

The Journal of Organic Chemistry

ARTICLE

Scheme 1. Expected Stereoselectivity Due to Steric Hindrance by the Ring Substituent in a LA Complexed Oxazolidinone and Mechanism Proposed by Evans et al.1g To Explain Their Experimental Results in DielsAlder Cycloadditions

Scheme 2. Rationalization of the Stereoselectivity in DA8 Reactions by Way of Lowest Energy Conformation (Antiparallel Carbonyl Groups) of the Amide Chain

Table 1. Yield and Diastereomeric Ratios (dr) Reported by Cardillo for the Conjugated Addition of O-Benzylhydroxylamine to 3 in the Presence of Different LAs2a,9

Scheme 3. Formation of Parallel Carbonyl Bicomplexes with BF3 3 Et2O As Proposed by Cardillo2a

CR-re face, while the isopropyl group in the auxiliary ring is too far away to directly interfere. To verify the applicability of our proposal to other important reactions, we decided to test our model8 on the system studied by Cardillo et al.2a,9 in the conjugated addition of O-benzylhydroxylamine to chiral R,β-unsaturated N-acylimidazolidinones 3 and 4 (Scheme 3), catalyzed by BF3, AlMe2Cl, and ZnCl2. This is a very interesting work that reports a broad range of complexing conditions, as BF3 should be able to form only open neutral complexes,10 whereas ZnCl2 can form neutral open or chelated complexes and AlMe2Cl can form neutral open complexes or charged chelated complexes. In this manuscript we show that the

entry

LA

equiv

yield (%)

dr (R:S)

1

BF3 3 Et2O

2

BF3 3 Et2O

1.0

80

90:10

2.0

>95

3

90:10

AlMe2Cl

1.0

65

74:26

4

AlMe2Cl

2.0

70

80:20

5

ZnCl2

1.0

92

45:55

selectivities obtained by Cardillo2a,9 can be fully rationalized by mechanisms based on open-chain antiparallel carbonyl complexes and that parallel carbonyl conformations, with or without chelate formation, induce no selectivity or inverted selectivity, respectively. These results can be extrapolated to other reactions based on similar auxiliaries and represent a good support to our previous study on DA reactions.8

’ RESULTS AND DISCUSSION The conjugated addition of O-benzylhydroxylamine to chiral R,β-unsaturated N-acylimidazolidinones (3 and 4) in the presence of BF3 3 Et2O,2a AlMe2Cl, or ZnCl29 was reported by Cardillo et al. (selected results in Table 1). The results in Table 1 are quite unexpected, as the chelating LAs (entries 35) should induce similar selectivities (excess of R isomer), whereas boron trifluoride (a nonchelating LA) is 6998

dx.doi.org/10.1021/jo201318f |J. Org. Chem. 2011, 76, 6997–7004


Figure 1. Relative free energies (ΔG) and dipole moments (μ), calculated in gas phase (roman) and solvent (italic), of four possible conformers of imidazolidinone 3. The larger differences calculated in gas phase between the parallel and antiparallel conformers are due to the greater dipole moments calculated in the parallel structures.

expected to induce the inverse selectivity (excess of S isomer) due to the hindrance of the aromatic ring in the auxiliary. The selectivity obtained with catalysis by AlMe2Cl was explained according to the Evans proposal, but the other selectivities are not easy to justify. Thus, in order to rationalize the selectivity obtained with BF3 3 Et2O and supported by NMR results that dismissed the formation of a pentacoordinated boron fluoride chelate, Cardillo suggested a parallel complex with coordination of each carbonyl group with a BF3 molecule (Scheme 3). According to Cardillo,2a this complex would be formed due to electrostatic attractions between the LA at the chain carbonyl group and the partial positive ring carbonyl moiety, leading to the unexpected observed selectivity (excess of R isomer). The NMR data presented by the authors to support this proposal are compatible with the formation of the parallel complex, but there is no data that unequivocally support its formation. An interesting observation, reported by the authors but that was not relevant in the discussion, is the fact that the diastereoselectivity is independent of the LA concentration. Thus, if the bicomplex is mandatory for catalysis, it implies that the monocomplex at the chain carbonyl group would be totally inefficient, a conclusion that contradicts other results, namely, those originally obtained by Evans for DA additions to R,β-unsaturated N-acyloxazolidinones.1g The possible conformers of compound 3 are shown in Figure 1. In the most stable conformation of compound 3, the double bond is in s-cis orientation and the two carbonyl groups adopt the antiparallel conformation (3CA). The relative energy values in Figure 1 are in agreement with previous results5 and

ARTICLE

indicate that the populations of the minor rotamers in the reactants have to be minimal. Cardillo’s results were obtained with AlMe2Cl, BF3, and ZnCl2.2a,9 As BF3 is not a good LA for modeling purposes, as it establishes very strong electrostatic contacts that are overestimated too much in the calculations, we study computationally only aluminum and zinc complexes, while our NMR studies focus on boron trifluoride and aluminum complexes. To simplify the discussion of our data, we calculated the reaction pathway not only via AlMe2Cl complexes but also via AlCl3 complexes. The advantage of using AlCl3 is that it originates a low number of rotamers, while from the electrostatic point of view, AlCl3 can be looked as a middle term between BF3 and AlMe2Cl, despite being a stronger LA. The results calculated with AlCl3, AlMe2Cl, or ZnCl2 are always given in the tables and in the Supporting Information, but the discussion is based on AlCl3 complexes. Three types of complexes with antiparallel carbonyl orientation have to be studied, as well as a bicomplex with the imidazolidinone in its parallel conformation and, for ZnCl2 and AlMe2Cl, chelated complexes in which the LAs connect simultaneously to the two carbonyl groups (Figure 2). Complexes with the s-trans conformation of the crotonyl moiety are not discussed, as their energies are too high to be relevant (see Supporting Information). According to Figure 2, in low concentrations of AlCl3 there is preferential complexation at the chain carbonyl group (MAC), as the monocomplex at the ring carbonyl group is more energetic by ca. 1.7 kcal mol1, while the extra stabilization resulting from the bicomplexation (19.0 kcal mol1 for BMA and 12.5 kcal mol1 for BMP) is lower than the first complexation at the chain carbonyl group (26.2 kcal mol1). Figure 2 also indicates that the parallel complex (BMP) has no tendency to be formed, as it is 6.5 kcal mol1 less stable than BMA. Thus, we can conclude that, according to our calculations, at low LA concentrations (1 equiv of LA) BMA will be formed, in accordance with the NMR results discussed below. From the data in Figure 2, we can also conclude that although AlMe2Cl behaves as AlCl3, ZnCl2 preferentially forms chelated complexes, as it is usually accepted. In most of the published reports, the characterization of the conformational behavior of free acylimidazolidinones or acyloxazolidinones and the modifications induced by the complexation with Lewis acids have been mainly addressed by the analysis of 1 H or 13C chemical shift changes, upon the addition of Lewis acids.2a,9,11 However, while 1H and 13C spectra indeed give useful information on complex formation, we have recently shown that, for related R,β-unsaturated N-acyloxazolidinones, the observed 1 H and 13C chemical shifts do not allow differentiation between chelates and bicomplexes.8 We have also demonstrated that, for the N-acyloxazolidinone system, substantial evidence for the formation of a chelate could be obtained in a nuclear Overhauser effect (NOE) based NMR experiment, since the rotation from the carbonyl antiparallel to the parallel conformation introduces dramatic changes in the interproton distances, which can be probed through the observation of NOE contacts. We have also shown that the complexation with Mg(ClO4)2 was a very efficient method of producing a chelate complex in solution and a straightforward way of recording NMR data of R,β-unsaturated N-acyloxazolidinone chelates with parallel carbonyl conformation. This information was then used as reference in studies with different Lewis acids.8 6999



ARTICLE

Figure 3. Comparison of the most stable conformations of compound 3, with antiparallel (3CA) or parallel (3CP) carbonyl conformations, and indication of the H20 /H4 interproton distances.

Figure 2. Relative free energies (ΔG), calculated in gas phase (roman) and solvent (italic), of five possible complexes (lowest energy conformers) between AlCl3, AlMe2Cl, or ZnCl2 and imidazolidinone 3. ZnCl2 bicomplexes are not considered, as the chelated complex is already 7.8 kcal mol1 more stable than the monocomplex MAC. Chelates with AlMe2Cl and AlCl3 are not shown because of their charged nature. Complexes with the s-trans conformation of the crotonyl moiety are not discussed, as their energies are too high to be relevant (around 4 kcal mol1 higher energy than the s-cis conformers; see Supporting Information).

For the present work, we have followed a similar approach in order to investigate the preference for the formation of chelates or open complexes with parallel or antiparallel carbonyl conformations, when BF3 or AlMe2Cl is added to compound 3. According to the data in Figure 1, the 3D structures of the most stable conformations of compound 3 in the antiparallel (3CA) or parallel (3CP) carbonyl orientations are those depicted in Figure 3. As in the parallel carbonyl conformation the distance between protons H20 /H4 is much shorter (less than 2.5 Å) than in the antiparallel conformation, the first should lead to strong NOE contacts, while the antiparallel conformation should have very weak or even no visible contacts. Following the above ideas, we have submitted compound 3 to a NOESY study, in the absence and in the presence of Mg(ClO4)2 (Figure 4), and the results are compared with the NOESY recorded in the presence of BF3 3 Et2O or AlMe2Cl, as discussed below. The NOESY spectrum of 3 obtained in the absence of LA (Figure 4A) shows no contact between protons H20 and H4, whereas in the presence of 1.0 equiv of Mg(ClO4)2 a strong contact can be observed between the same pair of protons (Figure 4B). Thus, in solution and in the absence of LA, compound 3 exists predominantly in conformation 3CA, whereas in the presence of Mg(ClO4)2 the conformation changes from anti- to parallel carbonyl orientation, due to chelate formation with the LA. The possibility of a fast exchange equilibrium between 3CP and

3CA in the absence of the LA was ruled out by performing low temperature experiments (from 25 to 80 °C) as, with the exception of the aromatic protons, the signals showed minimal temperature dependence (see Supporting Information). The downfield shift of H4 and the strong upfield shift of H20 , upon the addition of Mg(ClO4)2, are very similar to those previously observed with the related R,β-unsaturated N-acyloxazolidinone 2, under similar conditions.8 These shifts are indicative of complex formation and can be explained in terms of different carbonyl shielding environments due to rotation along the amide bond (see Supporting Information). In the sequence of the validation study described above, the titration of compound 3 with BF3 and AlMe2Cl was followed by 1 H NMR, leading to similar conclusions as previously obtained for the oxazolidinone system, when concerning the effect of complexation on chemical shifts (see Supporting Information).8 Our BF3 titration data in CD2Cl2 reproduces exactly the results previously reported by Cardillo.2a However, as we have demonstrated before,8 an analysis of the chemical shifts do not allow the differentiation between a chelate and a 2:1 complex (parallel or antiparallel), because in both cases the two carbonyl groups are coordinated and the electronic effects due to complexation or chelation are expected to be very similar. The fundamental difference among the various LAs is the number of equivalents of acid that are needed for total complexation of the initial imidazolidinone 3. Whereas full complexation is observed with 1.0 equiv of Mg(ClO4)2, 2.0 equiv of LA are needed when AlMe2Cl or BF3 is used. These results are consistent with the formation of chelates between compound 3 and Mg(ClO4)2 and open-chain mono- or bicomplexes with AlMe2Cl or BF3. The studies indicate also that for low LA concentrations, the first complexation takes place preferentially at the chain carbonyl group. The NMR results are in full accordance with the theoretical data. For a detailed analysis, see Supporting Information. Once the stoichiometry of the BF3 and AlMe2Cl complexes was established, the conformations were further investigated by NOESY experiments (Figure 5), and the conclusions are, in all aspects, identical to the results previously obtained for the complexation of compound 2 and AlMe2Cl.8 Despite the partial overlap of the H4/H20 resonances, the analysis of the NOESY spectra and the comparison with the Mg(ClO4)2 experiment (Figure 4B) reveals that, for BF3 and AlMe2Cl complexes, the carbonyl groups are in the antiparallel conformation, since no significant contact can be detected between H20 and H4 (Figure 5A1 and 5B1). 7000



ARTICLE

Figure 4. Expansion of the NOESY spectra of 3 obtained (A) in CD2Cl2 and (B) in the presence of 1.0 equiv of Mg(ClO4)2 in CD3CN.

Figure 5. Expansion of the NOESY spectra of 3 obtained in the presence of (A) 2.0 equiv of BF3 3 Et2O, with a detailed view of the H4/H20 region (A1), and (B) in the presence of 2.0 equiv of AlMe2Cl, with detailed view of the H4/H20 region (B1).

As the NMR confirms the theoretical data, indicating that BF3 and AlMe2Cl form open-chain carbonyl antiparallel complexes with 3 (monocomplexes preferentially at the chain carbonyl group and bicomplexes for concentrations of LA over 1 equiv) and Cardillo’s results2a,9 show that different concentrations of these LAs induce similar selectivities, an alternative mechanism to that proposed by Cardillo2a has to be established, based on carbonyl antiparallel complexes and able to account for the similar induction of selectivity in reactions via mono- or bicomplexes. Thus, we studied the applicability and performance of the mechanism recently proposed by us for DA reactions of derivatives of the Evans auxiliary catalyzed by aluminum-based LAs (Scheme 2),8 and the results are discussed below (Figure 6 and Table 2). In the reaction studied by Cardillo2a,9 O-benzylhydroxylamine was used as reagent. However, this amine offers special difficulties in modeling studies, as it allows for several possible conformers that strongly complicate the calculations and analysis of the results. Thus, we use dimethylamine (DMA) in the following discussion, as it is not much bulkier than O-benzylhydroxylamine and originates TSs with a single conformer. Due to the slightly

larger bulkiness of DMA and reduced electrostatic contacts, slightly overestimated calculated selectivities are obtained. As a result and aiming at an easy comparison between calculated and experimental values, we discuss the results using a scaling factor when the amine is DMA. However, Table 2 also shows the selectivities calculated without any scaling. Results obtained with methoxyamine (MA), an amine similar to O-benzylhydroxylamine via MAC- and BMA-AlCl3 complexes, are also shown in Table 2. These results are similar to those obtained with DMA, after application of the scaling factor (see Computational Methods).12 The results in Table 2 and Figure 6 indicate that our model can indeed explain the experimental selectivity, when the reaction occurs via the monocomplex at the chain carbonyl group or via the bicomplex in antiparallel carbonyl orientation. The analysis of the 3D models in Figure 6 indicate that, in order to minimize the steric interactions, the LA bound to the chain carbonyl group has to orientate itself away from the aromatic ring in the auxiliary, thus hindering the Cβ si-face more efficiently than the aromatic ring in the auxiliary hinders the Cβ re-face. The net result is the preferred attack at the Cβ re-face. 7001



ARTICLE

Table 2. Lowest Activation Free Energies (ΔG) and Diastereomeric Ratios (dr) for the Addition of DMA to Imidazolidinone 3 Complexed with AlCl3, AlMe3, and AlMe2Cla dr (R:S) entry

LA

complex

unscaled

scaled

1

R

40.8; 40.9 71.1:28.9 63.4:36.6

S

41.6; 41.4

AlCl3

MAC

4

R

30.1; 30.9 94.2:5.8

S

31.0; 32.5 34.5; 35.9 55.0:45.0 53.1:46.9 35.5; 36.0

5 6

MAR

R S

7

BMA

R

25.4; 27.2 93.5:6.5

S

26.8; 28.8

8 9

MAC MA

10 11

BMA MA

12

R

31.8; 34.8 88.0:12.0

S

31.9; 35.0

R

27.8; 30.6 72.5:27.5

S

27.2; 31.1 36.4; 35.6 98.7:1.3 38.3; 38.1

84.8:15.2

80.9:19.1

13 14

Al(CH3)3 MAC

R S

15

MAR

R

36.8; 37.8 56.2:43.8 53.8:46.2

S

37.7; 38.0

16 17

BMA

18 19

AlMe2Cl

MAC

20

93.4:6.6

R

33.1; 34.6 41.7:58.3 44.9:55.1

S

32.7; 34.4

R

33.4; 33.7 90.3:9.7

S

34.4; 34.9 36.5; 37.3 41.7:58.3 46.2:53.8 36.8; 37.0

21 22

MAR

R S

23

BMA

R

29.3; 30.5 92.4:7.6

S

30.1; 32.4

24

As Cardillo2a proposed a mechanism based on open-chain parallel carbonyl complexes, we decided to study also this possibility, even knowing that such complexes should not be formed, according to our theoretical data (Figure 2). We discuss this possibility using AlCl3, but the results for AlMe2Cl are also given in Table 3 and Figure 7. The results in Table 3 (entries 14) and Figure 7 indicate that bimetal parallel carbonyl complexes indeed catalyze the addition reaction, but they also clearly show that inverted selectivity would be obtained. This is well explained by our model, as the LA at the ring carbonyl group hinders the Cβ si-face and forces the LA at the chain carbonyl group to orientate into the opposite direction, laying at the Cβ re-face. Thus, the Cβ si-face becomes less hindered, which leads to the final product with (S) configuration. This result shows that not only does our proposed model totally explain the experimental selectivity obtained by Cardillo, but it also clearly shows that the model based on a carbonyl parallel bicomplex induces the opposite selectivity, which is a strong indication of its low probability. An interesting result from Cardillo9 was obtained with ZnCl2 as LA, a compound that is supposed to easily form neutral chelates.13 While our theoretical values indeed indicate that this LA should form chelated complexes with compound 3 (Figure 2),

ΔG

2 3

Figure 6. Unscaled relative free energies (ΔG) and scaled diastereomeric ratios (dr, in solvent) for the addition of DMA to imidazolidinone 3 in antiparallel carbonyl orientation, complexed with one or two molecules of AlCl3 or AlMe2Cl, calculated in gas phase (roman) and solvent (italic). Only the most stable structures are shown.

epimer

79.7:20.3

80.0:20.0

a

Calculated in gas phase (roman) and solvent (italic). The diastereomeric ratios were calculated from the unscaled and scaled PCM energy values. When the LA is AlMe2Cl, the selectivity via MAC or MAR was calculated by Boltzmann averaging over 6 possible TSs and 18 possible TSs for BMA. When the amine is methoxyamine (MA) 4 possible TSs were used for AlCl3. When more than two TSs are considered, the table lists the relative energies of the lowest R and S structures.

Cardillo’s results indicate that ZnCl2 induces no selectivity (Table 1). This observation strongly conflicts with Cardillo’s mechanistic proposal and needs to be tested against our own model. A positive result will be a very good confirmation of the validity of our model, as it will be able to justify all of the experimental results obtained by Cardillo and can also be easily extrapolated to other systems. We studied the catalysis using this LA, and the results are in Table 3 (entries 7 and 8) and Figure 8.14 The theoretical data in Table 3 and Figure 8 indicate that the chelated complex with ZnCl2 induces no stereoselectivity, in full accordance with the experimental results. This is not unexpected and agrees with recent results published by us15 on the induction of selectivity in radical additions. As the β carbon atom is too far away from the auxiliary substituent, the attacking amine does not suffer a relevant steric contact, even when the amide bond conformation is inverted to form the chelate. According to the data in Table 3 (entries 5 and 6), a similar result is obtained when the chelating LA is AlMe2Cl.16 The above-discussed results clearly indicate that the conjugated addition of amines to chiral R,β-unsaturated N-acylimidazolidinones catalyzed by AlMe2Cl or BF3 does not proceed 7002



ARTICLE

Table 3. Lowest Activation Free Energies (ΔG) and Scaled and Unscaled Diastereomeric Ratios (dr, in Solvent) for the Addition of DMA to Imidazolidinone 3 in Carbonyl Parallel Conformation, Complexed with AlCl3, AlMe2Cl, and ZnCl2a dr (R:S) entry 1

LA AlCl3 AlMe2Cl

ZnCl2

R

20.5; 26.1

4.4:95.6

13.1:86.9

S

19.7; 24.3 26.6:73.4

34.3:65.7

52.8:47.2

51.7:48.3

60.1:39.9

56.2:43.8

R

30.0; 30.8

S

29.0; 30.4

MCP

R

24.7; 28.1

S

25.1; 28.2

MCP

R S

32.7; 33.6 31.9; 33.8

BMP

6 7 8

scaled

BMP

4 5

unscaled

epimer

2 3

ΔG

complex

a

Calculated in gas phase (roman) and solvent (italic). When the LA is AlMe2Cl, the selectivity via BMP was calculated by Boltzmann averaging over 9 possible TSs.

chiral R,β-unsaturated N-acyloxazolidinones catalyzed by AlMe2 Cl. The conclusion that the traditional concept of direct steric hindrance by the substituents in oxazolidinone or imidazolidinone auxiliaries does not have to be always correct gets support on a recent paper by Hsung,17 where the unexpected selectivity obtained in (4 + 3) cycloaddition reactions of oxyallyls was explained not as a result of steric contacts but of CH-π interactions between the attacking cyclopentadiene and the aromatic ring in the auxiliary.

’ CONCLUSIONS In this paper we studied, by NMR and theoretical methods, the catalyzed conjugated addition of amines to acylimidazolidinones that was reported by Cardillo. Our results clearly show that aluminum chelates are not involved in the reaction, as they induce no selectivity, while Cardillo observed high experimental selectivities. Our data equally show that bicomplexes with carbonyl parallel orientation, proposed by Cardillo to justify the experimental selectivity with nonchelating Lewis acids, indeed induce the opposite selectivity and have also to be dismissed. On the other hand, our mechanistic proposal, based on open-chain carbonyl antiparallel complexes, allows for the full rationalization of the data obtained by Cardillo with aluminum, boron, or zinc Lewis acids and supports our previous proposal on DA cycloadditions of dienes to Evans chiral auxiliary derivatives. Extrapolation to other systems is possible, which makes this proposal a very powerful tool to help in the rationalization of known and new obtained data. ’ COMPUTATIONAL METHODS

Figure 7. Unscaled relative free energies (ΔG) and scaled diastereomeric ratios (dr, in solvent) for the addition of DMA to imidazolidinone 3 in parallel carbonyl orientation, complexed with two molecules of AlCl3 or AlMe2Cl, calculated in gas phase (roman) and solvent (italic). Only the most stable structures are shown.

Full geometry optimizations have been performed with the Gaussian 03, Revision E.01, software package18 suite of programs employing density functional theory (DFT)19 with the hybrid functional PBE1PBE20 and the 6-31G(d) basis set. Harmonic vibrational frequencies have been calculated for all located stationary structures to verify whether they are minima or transition states. Zero-point energies and thermal corrections have been taken from unscaled vibrational frequencies. Free energies of activation, unless otherwise stated, are given at 25 °C. The energy values have been refined by single point DFT calculations at the PBE1PBE/ 6-311+G(2d,p) level of theory, over the optimized gas phase geometries. Solvent effects in dichloromethane (DCM) have been taken into account by single point calculations with the polarizable continuum model (PCM)21 over the respective gas-phase geometries. Complexation and activation energies are calculated relative to the reagents. As the selectivities obtained when the reacting amine is dimethylamine are, as expected, slightly overestimated, scaled values are also shown in the tables. A scaling factor of 0.615 was used over the obtained PCM computed activation free energy differences. This value was obtained by fitting the theoretical and experimental diastereomeric ratios obtained for the reaction catalyzed by 2.0 equiv of AlMe2Cl. All bond lengths are in angstroms (Å), energies are in kcal mol1, and dipole moments are in debye (D).

Figure 8. Unscaled relative free energies (ΔG) and diastereomeric ratios (dr, in solvent) for the addition of DMA to imidazolidinone 3 as a chelated complex with ZnCl2 (top) or AlMe2Cl (bottom), calculated in gas phase (roman) and solvent (italic). The diastereomeric ratio was calculated from the scaled PCM energy values.

’ EXPERIMENTAL SECTION

via chelated complexes, as the selectivity would be very low, or via open-chain carbonyl parallel complexes, as the selectivity would be inverted. These conclusions can be also applied to other systems, under similar reaction conditions. In particular, the results here obtained are a very strong support for the alternative mechanism proposed by us8 to justify the selectivity obtained by Evans1g in DA additions of dienes to

NMR Methods. 4S,5R-3 was prepared according to reported procedures.9 All NMR spectra were recorded on a NMR spectrometer, operating at 400.13 MHz for hydrogen. Chemical shifts were referenced to the chemical shift of the residual proton of the solvent peak. Standard 2D homonuclear (COSY) and heterenuclear correlations (HMQC) were performed whenever necessary for spectral assignment. The twodimensional gs-NOESY experiments were acquired in phase-sensitive mode, using standard pulse sequences22 with mixing times ranging from 7003


The Journal of Organic Chemistry 500 ms up to 1.5 s. Generally, 8 or 16 scans and 256 F1 increments were obtained, and the spectral width in both dimensions was 5000 Hz.

’ ASSOCIATED CONTENT

bS

Supporting Information. Extra structures for the uncatalyzed conjugated addition with respective energies. Cartesian coordinate matrixes and electronic energies of all calculated structures. Extra NMR spectral information. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected].

’ ACKNOWLEDGMENT We are grateful to Fundac-~ao para a Ciência e Tecnologia (SFRH/BPD/26676/2006, PTDC/QUI-QUI/104056/2008 and Projecto de Re-equipamento Científico, 2005 - Computer Cluster) for financial support. The NMR spectrometers are part of the National NMR Network (RNRMN) and are funded by Fundac-~ao para a Ciência e a Tecnologia (FCT). ’ REFERENCES (1) (a) Evans, D. A.; Ennis, M. D.; Mathre, D. J. J. Am. Chem. Soc. 1982, 104, 1737–1739. (b) Evans, D. A. Aldrichimica Acta 1982, 15, 23–32. (c) Evans, D. A.; Morrissey, M. M.; Dorow, R. L. J. Am. Chem. Soc. 1985, 107, 4346–4348. (d) Evans, D. A.; Bartroli, J.; Shih, T. L. J. Am. Chem. Soc. 1981, 103, 2127–2129. (e) Abdel-Magid, A.; Pridgen, L. N.; Eggleston, D. S.; Lantos, I. J. Am. Chem. Soc. 1986, 108, 4595–4602. (f) Evans, D. A.; Weber, A. E. J. Am. Chem. Soc. 1986, 108, 6757–6761. (g) Evans, D. A.; Chapman, K. T.; Bisaha, J. J. Am. Chem. Soc. 1988, 110, 1238–1256. (2) (a) Cardillo, G.; Gentilucci, L.; Gianotti, M.; Tolomelli, A. Org. Lett. 2001, 3, 1165–1167. (b) Nicolas, E.; Russell, K. C.; Knollenberg, J.; Hruby, V. J. J. Org. Chem. 1993, 58, 7565–7571. (c) Li, G.; Jarosinski, M. A.; Hruby, V. J. Tetrahedron Lett. 1993, 34, 2561–2564. (d) Hale, K. J.; Cai., J.; Delisser, V.; Manaviazar, S.; Peak, S. A.; Bhatia, G. S.; Collins, T. C.; Jogiya, N. Tetrahedron 1996, 52, 1047–1068. (3) Gothelf, K. V.; Hazell, R. G.; Jorgensen, K. A. J. Am. Chem. Soc. 1995, 117, 4435–4436. (4) (a) Caddick, S.; Jenkins, K.; Treweeke, N.; Candeias, S. X.; Afonso, C. A. Tetrahedron Lett. 1998, 39, 2203–2206. (b) Schreiner, P. R.; Wittkopp, A. Org. Lett. 2002, 4, 217–220. (c) Yamamoto, Y.; Onuki, S.; Yumoto, M.; Asao, N. J. Am. Chem. Soc. 1994, 116, 421–422. (5) (a) Santos, A. G.; Candeias, S. X.; Afonso, C. A. M.; Jenkins, K.; Caddick, S.; Treweeke, N. R.; Pardoe, D. Tetrahedron 2001, 57, 6607–6614. (b) Santos, A. G.; Pereira, J.; Afonso, C. A. M.; Frenking, G. Chem.—Eur. J. 2005, 11, 330–343. (6) (a) Caddick, S.; Jenkins, K. Tetrahedron Lett. 1996, 37, 1301–1304. (b) Kubota, H.; Kubo, A.; Nunami, K. Tetrahedron Lett. 1994, 35, 3107–3110. (c) Nunami, K.; Kubota, H.; Kubo, A. Tetrahedron Lett. 1994, 35, 8639–8642. (d) Kubota, H.; Kubo, A.; Takahashi, M.; Shimizu, R.; Da-te, T.; Okamura, K.; Nunami, K. J. Org. Chem. 1995, 60, 6776–6784. (e) Kubo, A.; Takahashi, M.; Kubota, H.; Nunami, K. Tetrahedron Lett. 1995, 36, 6251–6252. (f) Kubo, A.; Kubota, H.; Takahashi, M.; K. Nunami, K. Tetrahedron Lett. 1996, 37, 4957–4960. (g) Kubo, A.; Kubota, H.; Takahashi, M.; Nunami, K. J. Org. Chem. 1997, 62, 5830–5837. (7) Caddick, S.; Afonso, C. A. M.; Candeias, S. X.; Hitchcock, P. B.; Jenkins, K.; Murtagh, L.; Pardoe, D.; Santos, A. G.; Treweeke, N. R.; Weaving, R Tetrahedron 2001, 57, 6589–6605.

ARTICLE

(8) (a) Bakalova, S. M.; Santos, A. G. J. Org. Chem. 2004, 69, 8475–8481. (b) Bakalova, S. M.; Santos, A. G. Eur. J. Org. Chem. 2006, 1779–1789. (c) Bakalova, S. M.; Duarte, F. J. S.; Georgieva, M. K.; Cabrita, E. J.; Santos, A. G. Chem.—Eur. J. 2009, 15, 7665–7677. (9) Amoroso, R.; Cardillo, G.; Sabatino, P.; Tomasini, C.; Trere, A. J. Org. Chem. 1993, 58, 5615–5619. (10) (a) Hawkins, J. M.; Loren, S. J. Am. Chem. Soc. 1991, 113, 7794–7795. (b) Hawkins, J. M.; Loren, S.; Nambu, M. J. Am. Chem. Soc. 1994, 116, 1657–1660. (11) (a) Laszlo, P.; Teston, M. J. Am. Chem. Soc. 1990, 112, 8750–8754. (b) Castellino, S.; Dwight, W. J. J. Am. Chem. Soc. 1993, 115, 2986–2987. (c) Childs, R. F.; Mulholland, D. L.; Nixon, A. Can. J. Chem. 1982, 60, 801–808. (12) A second order mechanism in the amine can be considered, which leads to a six-membered ring TS instead of the four-membered ring we discuss. However, our calculations indicate that such mechanism does not allow for the rationalization of the reaction selectivity, as all calculated complexes originate very low selectivities. Thus, while the matrixes are available in Supporting Information, no discussion is made in the main text. (13) Domingo, L. R.; Andres, J.; Alves, C. N. Eur. J. Org. Chem. 2002, 2557–2564. (14) According to the complexation energies in Figure 2, ZnCl2 forms preferentially the chelated complex, as its energy is lower than those of the monocomplexes. Thus, we only discuss the reaction via the chelated complex. (15) Georgieva, M. K.; Duarte, F. J. S.; Bakalova, S. M.; Santos, A. G. Eur. J. Org. Chem. 2010, 4841–4850. (16) In the case of DA reactions1g,8 in which the reacting species is large and the reaction occurs simultaneously at carbons R and ss, the result is different, as the rotation of the amide-bond indeed brings the attacking reagent near the auxiliary substituent, thus inducing good selectivity.8 (17) Krenske, E. H.; Houk, K. N.; Lohse, A. G.; Antoline, J. E.; Hsung, R. P. Chem. Sci. 2010, 1, 387–392. (18) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A.; , Jr., Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; M. A. Al-Laham, Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; , Pople, J. A. Gaussian 03, Revision E.01; Gaussian, Inc.: Wallingford, CT, 2004. (19) (a) Parr, R. G.; Yang, W. Density Functional Theory of Atoms and Molecules; Oxford University Press: Oxford, 1989. (b) Koch, W.; Holthausen, M. C. A Chemist’s Guide to DFT, 2nd ed.; Wiley: New York, 2001. (20) (a) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1996, 77, 3865–3868. (b) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1997, 78, 1396. (c) Adamo, C.; Barone, V. J. Chem. Phys. 1999, 110, 6158–6169. (d) Wheeler, S. E.; Moran, A.; Pieniazek, S. N.; Houk, K. N. J. Phys. Chem. A 2009, 113, 10376–10384. (21) Tomasi, J.; Mennucci, B.; Cammi, R. Chem. Rev. 2005, 105, 2999–3093. (22) Berger, S.; Braun, S. 200 and More NMR ExperimentsA Pratical Course; Wiley-VCH: Weinheim, 2004.

7004


Lewis Acid Catalyzed Reactions of Chiral Imidazolidinones and

Recommend Documents